- Email: [email protected]
SiC composites in liquid fluoride salt environment
Accepted Manuscript Title: Microstructure evolution and high-temperature mechanical properties of SiCf /SiC composites in liquid fluoride salt environment Authors: Hongda Wang, Qian Feng, Zhen Wang, Haijun Zhou, Yanmei Kan, Jianbao Hu, Shaoming Dong PII: DOI: Reference:
S0010-938X(16)30734-X http://dx.doi.org/doi:10.1016/j.corsci.2017.05.016 CS 7092
To appear in: Received date: Revised date: Accepted date:
10-9-2016 16-5-2017 23-5-2017
Please cite this article as: Hongda Wang, Qian Feng, Zhen Wang, Haijun Zhou, Yanmei Kan, Jianbao Hu, Shaoming Dong, Microstructure evolution and high-temperature mechanical properties of SiCf/SiC composites in liquid fluoride salt environment, Corrosion Sciencehttp://dx.doi.org/10.1016/j.corsci.2017.05.016 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Microstructure evolution and high-temperature mechanical properties of SiCf/SiC composites in liquid fluoride salt environment Hongda Wanga,b,c, Qian Fengd, Zhen Wanga,b,, Haijun Zhoua,b,*, Yanmei Kana,b, Jianbao Hua,b, Shaoming Donga,b,* a
State Key Laboratory of High Performance Ceramics and Superfine Microstructure, Shanghai Institute of
Ceramics, Chinese Academy of Sciences, Shanghai 200050, China. b
Structural Ceramics and Composites Engineering Research Center, Shanghai Institute of Ceramics, Chinese
Academy of Sciences, Shanghai 200050, China. c
University of Chinese Academy of Sciences, Beijing 100049, China.
d
Analysis and Testing Center, Donghua University, Shanghai 201600, China.
*Corresponding author: Haijun Zhou: Tel.: +86 21 69906076
Fax: +86 21 69906085
E-mail: [email protected] Address: Shanghai Institute of Ceramics, Chinese Academy of Sciences, No. 1295, Dingxi Road, Changning District, Shanghai 200050, P. R. China. Shaoming Dong: Tel.: +86 21 52414324
Fax: +86 21 69906085
E-mail: [email protected] Address: Shanghai Institute of Ceramics, Chinese Academy of Sciences, No. 1295, Dingxi Road, Changning District, Shanghai 200050, P. R. China.
1
Highlights:
Corrosion behavior of SiCf/SiC CMCs in fluoride salt was studied.
O-contained boundaries result in the non-uniform corrosion of SiC matrix.
The Si-F coordinate bond plays important roles in SiC corrosion.
Corrosion damage of matrix leads to degradation of mechanical properties.
Abstract High temperature mechanical properties and microstructure evolution of chemical-vapor-infiltrated SiC fiber reinforced SiC ceramic matrix composites (SiCf/SiC) in 46.5LiF-11.5NaF-42.0KF (mol. %, FLiNaK) eutectic salt were investigated. The results indicate that the corrosion of SiCf/SiC composites was accelerated with increase of corrosion temperature. Interlayer boundaries in SiC matrix with higher oxygen content were corroded preferentially. The F- formed Si-F coordinate bond with SiC and replaced carbon, followed by the corrosion of SiC matrix. While single SiC fiber tensile strength suffered no depravation after corrosion, corrosion and damage of SiC matrix led to deteriorated high temperature mechanical properties of SiCf/SiC composites.
Key words: A. Ceramic matrix composites; A. Molten salt; B. SEM; B. XRD; C. Exfoliation corrosion; C. High temperature corrosion
1. Introduction Silicon carbide (SiC) continuous fiber-reinforced SiC matrix composites (SiCf/SiC) have a multitude of excellent properties, including high temperature mechanical performance with better mechanical damage tolerance over inherently brittle monolithic SiC[1, 2], outstanding thermal conductivity, low activation to 2
neutron-induced radioactivity[3-7], which make the SiCf/SiC composites a promising candidate material for structural and functional components in nuclear reactors. Particularly, SiCf/SiC composites have nice compatibility with liquid fluoride salt (e.g. 46.5 mol % LiF- 11.5 mol % NaF-42.0 mol % KF eutectic molten salt, i.e. FLiNaK molten salt). This unique properties of SiCf/SiC composites make it achieve wide attention due to their potential application in heat-exchanger, control rod and other structure components in Molten Salt Reactor (MSR)[8-11]. However, SiC undergoes corrosion in liquid fluoride salt at high temperature because of the disequilibrium in thermodynamics between SiC and fluoride salt[11]. Obviously, the corrosion of SiC compromises the advantages of SiCf/SiC composites in nuclear application. The corrosion behavior of SiCf/SiC composites in liquid fluoride salt has been seldom investigated up to now, especially the evolution of microstructure and mechanical performance. It is indicated that the impurities in fluoride salt and the composition of SiC-based materials lead to the incompatibility of SiCf/SiC composites to fluoride salt. The foreign ions (i.e. Ni2+, Cr2+ and Cr3+), which could be introduced during production of fluoride salt or by corrosion of Hastelloy N alloy, can accelerate the corrosion of SiC[12, 13]. Although the reason of optionally corrosion of SiC is still unknown, qualitative explanation with the respect of the standard Gibbs free energy of fluoride formation per F2 molecule of the Si and SiC phases has been widely accepted[14]. The Si-O bond in O-contained SiC phase would be attacked by F- to form Si-F bond, indicating that the oxygen in SiC accelerates the corrosion of SiC in fluoride salt[15]. Therefore, high-purity SiC is needed in nuclear reactor application. There is a variety of technology for preparing SiCf/SiC composites, such as nano-infiltration and transient eutectic phase[16], polymer impregnation and pyrolysis[17], reaction sintering[18] and chemical vapor infiltration (CVI) process[19]. Among these methods, the CVI process could form highly crystalline, nearstoichiometric SiC matrix and minimize damage of fibers during material fabrication, which are beneficial to 3
produce the nuclear grade SiCf/SiC composites[20]. However, few research has focused on the corrosion behavior of CVI SiCf/SiC composites and mechanical performance in liquid fluoride salt environment. In this work, the static corrosion behavior of CVI SiCf/SiC composites in molten FLiNaK salt at various temperature was studied. The microstructure evolution during the corrosion process, and the high temperature mechanical properties were investigated. 2. Experimental The materials used to evaluate corrosion behavior are SiC matrix composites reinforced by two-dimensional woven fabrics of KD-II SiC fibers (National University of Defense Technology, Changsha, China). Using the Methyltrichlorosilane (MTS) as the raw material and H2 as the carrier gas, the SiCf/SiC composites were prepared via CVI processing. Prior to the CVI process, a pyrolytic carbon (PyC) monolayer of ~ 100 nm in thickness was deposited on the SiC fiber surface through chemical vapor deposition. The fiber volume fraction, porosity and density of the as-received composites were ~ 40%, ~ 5.01 % and 2.74 g∙cm-3, respectively. The monolithic composites were machined into 35 × 4 × 2 mm3 bar for the corrosion experiment. Each test group contained 5 bars. According to the design goals of Generation IV reactors, the outlet temperature of MSR would be 700 °C ~ 800 °C[21], and the peak temperature of Advanced High Temperature Reactor would be up to 1000 °C[11]. To evaluate the performance of SiCf/SiC composites in the similar environment in reactor, 800 °C, 900 °C and 1000 °C were chosen as the corrosion temperature, and the corrosion time was 500 h[22]. The SiCf/SiC composites samples were transferred into a glove box under the protection of high purity argon gas. The atmosphere in the glove box was controlled at O2 < 0.1 ppm and moisture < 0.1 ppm all along. Before corrosion, the graphite crucible and SiCf/SiC samples were heat-treated in furnace in the glove box at 700 °C for 10 h to eliminate possible adsorption of air and other gases. Subsequently, SiCf/SiC samples were fixed on a graphite 4
support and dipped into ~ 80 g 46.5LiF-11.5NaF-42.0KF (mol. %, FLiNaK) fluoride eutectic molten salt (Shanghai Institute of Organic Chemistry, Chinese Academy of Sciences, Shanghai) in the graphite crucibles. The same SiC fibers used in SiCf/SiC composites were also introduced into the crucibles at the same time to study the effect of corrosion on mechanical properties of SiC fibers. And then, the crucibles were heated to 800 °C, 900 °C or 1000 °C for 500 h, respectively. After the high temperature corrosion process in fluoride eutectic molten salt, the SiCf/SiC composite samples were picked out from the crucible and immersed into 1 mol∙L-1 Al(NO3)3 solution to remove the fluoride salt solid adhered on the exterior and interior surfaces of the sample[23]. Finally, the samples were washed with deionized water using an ultrasonic cleaner, and then dried at 120 °C in a vacuum drying chamber. The samples were weighed by precision balance (BT-25-S, 0.01 mg, Sartorius) before and after the corrosion test, and the mass loss ratio was calculated to characterize the corrosion resistance of the composites. Every bar in each test group was measured 3 times independently to obtain the mass loss ratio and mass loss rate of each bar. Then data of 5 bars in each experiment group was used to calculate the average mass loss ratio, mass loss rate and the corresponding standard deviation. The mass loss ratio (ml) could be calculated by the formula below:
ml
m0 m 100% m0
(1)
where m0 and m are the mass of sample before and after corrosion, respectively. The apparent mass loss rate (v) was introduced to semi-quantitatively describe the corrosion rate:
m0 m m0 m S t 2( LH LW HW ) t
(2)
where S is surface area of sample, t is corrosion time, and L, W, H are the length, width, height of sample, respectively. Phase compositions of the samples were analyzed by X-ray diffraction (XRD, D8 Advance, 5
Bruker AXS Co. Ltd., GER) at room temperature, using Cu Kα radiation (λ = 0.154 nm) 2θ over the angles range of 10 ° ~ 80 °. High temperature flexural strength of the samples after corrosion was measured through a three-point-bending test (Instron-5500R, Instron Corp., Canton, USA) with a span of 30 mm and a loading rate of 0.5 mm/min in air atmosphere. For a comparison, the high temperature flexural strength of samples without corrosion was also measured. 3 bars were tested for each group. Polished cross-section and surface morphologies of the samples before and after corrosion at varying temperatures were characterized by scanning electron microscope (SEM, Quanta-250, FEI, USA) along with energy dispersive spectrometer (EDS, Oxford, UK). Additionally, the single fiber tensile strength of SiC fibers before and after corrosion was tested by single fiber electronic tensile strength tester (YG-001A, Hongda Fangyuan Electric Co. Ltd., China), using a span distance of 20 mm and stretching rate of 15 mm·min-1. 3. Results 3.1. Morphologies of SiCf/SiC composites The surface morphologies of as-received and corroded SiCf/SiC composites samples are given in Fig. 1. In comparison with the planar surface of untreated samples, rougher surface is observed after corrosion. After exposed in liquid FLiNaK salt at 800 °C for 500 h, matrix corrosion and exfoliation occurs. At higher temperature of 900 °C and 1000 °C, more intensive destruction, including extensive matrix layer exfoliation around fibers and fiber damage, was observed clearly. Polished cross-section microstructures of as-received and corroded samples at varying temperatures are shown in Fig. 2. Matrix exhibits typical feature of layer structure, which is resulted from repetition of CVI process. Also, residual pores can be observed in fiber bundle (Fig. 2a). Line scan result given in white box in Fig. 2a shows that interlayer boundaries contain slightly more oxygen than matrix. After corrosion in liquid fluoride salt at varying temperatures, the interlayer boundaries between deposited matrix layers are selectively 6
corroded, which leads to gaps formed between deposited layers. When corroded at 800 °C for 500 h, parts of interlayer boundaries and matrix around residual pores are eroded (pointed by arrows in Fig. 2b). After corrosion at 900 °C, interlayer boundaries between deposited matrix layers are almost completely corroded. Layer exfoliation is observed around the interior surface of residual pores. Corrosion at 1000 °C causes intensive damage of matrix, especially the serious layer exfoliation of deposited matrix layers. Mechanical damage of matrix around pores and corroded boundaries during polishing process is also found because of loose structure of matrix caused by corrosion. Owing to the formation of gaps in matrix resulted from the selective corrosion of interlayer boundaries, the molten salt would penetrate into the composites, leading to extension of corrosion in both exterior and interior of the sample. The microstructure evolution before and after corrosion indicates a non-uniform corrosion, which was different from the uniform corrosion of CVD SiC in fluoride salt reported in references [12, 15]. 3.2. Mass change Considering the non-uniform corrosion process (especially the optional corrosion of interlayer boundaries) caused by the anisotropic microstructure and heterogeneous composition of the composites, the normal expression on corrosion by using a mass loss per unit area or a thickness may not be appropriate. Therefore, the mass loss ratio is used to describe the corrosion damage degree of samples. At the same time, the apparent mass loss rate is employed to characterize the corrosion rate semi-quantitatively because of the change of corrosion surface area during corrosion. Table 1 shows the mass loss ratio and apparent mass loss rate of samples corroded at 800 °C, 900 °C and 1000 °C. As the corrosion temperature rises from 800 °C to 1000 °C, the mass loss ratio increases from 5.914 wt. % to 31.062 wt. %, while the apparent mass loss rate raises from 0.211 μg·mm-2·h-1 to 1.107 μg·mm-2·h-1, which suggests that higher corrosion temperature would aggravate the damage of SiCf/SiC composites in liquid fluoride salt. 7
3.3. X-ray diffraction pattern Fig. 3 shows the XRD patterns of the SiCf/SiC composites before and after corrosion at selected temperatures for 500 h. Different from the XRD pattern of as-received SiCf/SiC composites, diffraction peaks corresponding to Graphite-2H are detected in XRD patterns of the corroded samples. Peak intensity of the graphite phase increases with the corrosion temperature. Graphite-2H is believed to be the corrosion product of β-SiC in the molten salt. 3.4. High temperature mechanical property High temperature mechanical properties of SiCf/SiC composites under corrosive environment are important and fundamental for the application in salt-cooled reactors. Three-point bending test at corresponding corrosion temperature was employed to characterize the effect of corrosion in liquid fluoride salt on mechanical performance of SiCf/SiC composites. The flexural strength of samples before and after corrosion at varying temperatures are listed in Table 2. Compared with the flexural strength of as-received samples tested at corresponding corrosion temperatures, the flexural strength of samples after corrosion exhibits significant decrement. Flexural strength retention ratio of the samples after corrosion at 800 °C, 900 °C and 1000 °C compared to that of the as-received samples at corresponding temperature were 51.64 %, 65.26 % and 25.13, respectively. Both the flexural strength before and after corrosion decreases with the increase of the testing temperature, and the corroded samples suffer larger drop. Table 3 gives the porosity of each group samples before and after corrosion. Corrosion process leads to higher porosity, and larger sized pores are obtained at higher corrosion temperature. The high temperature stress - strain curves of the samples before and after corrosion are presented in Fig. 4. All curves contain the extended regions after the initial failure, which indicate the ductile rupture behavior for all the samples before and after corrosion. 8
Fig. 5 shows fracture surface morphologies of the SiCf/SiC composites. Long pull-out fibers are observed on the fracture surface of blank samples and corroded samples at 800 °C, while long pull-out fiber bundle instead of common pull-out fibers occurs at 900 °C and fiber bundle breaks with short finite length at 1000 °C. This is typical of crack arresting, deflecting and branching behavior which leads to the ductile rupture mode of the composites. Interestingly, compared with the brittle fracture matrix around the fiber bundle in blank sample, the corroded sample at 800 °C shows step-like morphology on fracture surface, which is similar to the fracture morphology of (SiC/PyC) multilayer[24]. 4. Discussion 4.1. Corrosion behavior The corrosion of SiC in fluoride salt is very complex, and the mechanism is still unknown. It is well known that the oxygen in SiC as well as the impurities in fluoride salt shows great influence on the corrosion behavior [12, 13, 15]. Compared with the CVD SiC, the SiCf/SiC composites with different microstructure may show distinctive corrosion behavior. As shown in Fig. 2, the appearance of gaps resulted from the preferential corrosion of interlayer boundaries indicates that the corrosion of SiCf/SiC composites would be non-uniform corrosion. EDS linear scan results shown in Fig. 2a demonstrates that oxygen content in interlayer boundaries is slightly higher than that in matrix. This may be associated with the repetition of densification process. To decrease the porosity of SiCf/SiC composites, component surface usually needs to be grinded to reopen the pores that sealed off during deposition, which guarantees the reactants to penetrate into the component during the subsequent infiltration process. However, oxygen absorption and slow oxidation would occur on the surface of deposited SiC layer when reopening the pores. The O-rich surface can be covered by subsequent deposited SiC layer, forming interlayer boundaries with high oxygen content between deposited SiC layers. Higher oxygen content leads to 9
worse corrosion resistance of interlayer boundaries than near-stoichiometric SiC[15]. When dipped into FLiNaK salt, Si-O bond in these interlayers would be attacked by F- to form more stable Si-F bond[25]. SiF4, which is believed to be the corrosion product, is considered to dissolve in FLiNaK salt and form SiF62-. As a result, the interlayer boundaries would be corroded into gaps, which destructs the bonding force between deposited layers and causes the layer exfoliation (Fig. 2c). At the same time, liquid salt infiltrates into the matrix through the corrosion gaps and further damages the matrix inside the composites (Fig. 2d). The XRD patterns shown in Fig. 3 also suggests that the corrosion of SiC in fluoride salt would be the loss of Si in SiC, with graphite as the corrosion product. The F ion has large electronegativity and is a kind of strong nucleophilic reagent. The outermost electron shell of Si atom has 4 electrons, which would form covalent bond with valence electrons of C atom in SiC through sp3 hybrid. However, there are still empty electronic orbits, d orbitals, in outermost electron shell of Si. During the corrosion, F- would provide common electrons to occupy the empty d orbitals of Si to form Si-F coordinate bond. This Si-F bond could change the symmetrical structure of [SiC4] tetrahedron and is conducive to further replacement of C by F. As a result, SiC bond is replaced by Si-F bond, and hybridization form of Si shifts from sp3 to sp3d2. The Si bonds with F in salt to form [SiF6]2-, which dissolved into FLiNaK salt[15]. The [SiF6]2- could decompose into SiF4 gas and escape out from liquid salt at high temperature, which contributes to the further corrosion and dissolution of SiC. At the same time, the C4- in SiC would be oxidized into C by impurities, such as NO3-[26]. And then, newly formed C may crystallizes into graphite, which might be similar to the graphitization process on the surface of SiC[27, 28]. After the dissolution of Si in form of [SiF6]2-, two or three successive carbon layers collapse into one layer to form graphite[29]. From Fig. 1 and Fig. 2, heavier corrosion damage of SiCf/SiC composites is observed at higher temperature, indicating that higher temperature would accelerate the corrosion of SiCf/SiC composites. This is coincident 10
with the mass loss ratio in Table 1, the XRD patterns in Fig. 3 obtained at selected temperatures and the porosity of samples shown in Table 3. According to the Arrhenius equation[30], higher temperature supplies enough energy for overcoming the reaction barrier, known as the activation energy (E a), and more effective collision would happen during reaction. The corrosion reaction would be accelerated at higher temperature. Consequently, this causes worse damage and larger mass loss ratio at higher corrosion temperature. 4.2 Mechanical properties Considering the microstructure evolution of samples corroded at varying temperature shown in Fig. 1 and Fig. 2, the serious decrement of high temperature flexural strength with increase of the corrosion temperature would be attributed to the corrosion damage caused by fluoride salt. Initially, the selective corrosion of O-contained interlayer boundaries destroys the integrity of matrix and lowers the capacity of matrix. As shown in Fig. 4, the slop of ascending portion in stress - stain curves of corroded samples decreases with increasing corrosion temperature, and is also smaller than that of as-received samples obtained at corresponding temperature. This suggests that the elastic modulus of samples degrades in corrosion process. Gaps resulted from the selective corrosion of interlayer boundaries should be responsible for this. As shown in Table 1, the appearance of gaps in matrix after corrosion increases the porosity in the composites. The relationship between elastic modulus and porosity could be expressed as the empirical equation below[31]:
E E0 (1 1.9P 0.9P2 )
(3)
where P is the porosity of material, E is the elastic modulus of material when the porosity is P, and E0 is the elastic modulus of fully dense material. After corrosion, higher the corrosion temperature is, larger porosity the sample would obtain, which lowers the modulus significantly. As a result, matrix load capacity would decrease and mechanical performance is deteriorated. 11
Secondly, the oxidation of PyC interface in SiCf/SiC composites during test may partly account for the degradation of mechanical properties. The high temperature flexural strength test was conducted in air atmosphere. The PyC interface would react with O2 during heating, and the oxidation rate of PyC increases with increased temperature. The damage of PyC interface would hinder the load transfer between SiC matrix and SiC fibers. In consequence, the flexural strength would decrease at higher temperature. Additionally, corrosion experiments of the same SiC fibers used in composites were conducted. Single fiber tensile strength of the SiC fibers is 1.72 GPa before corrosion and is 1.65 GPa, 1.73 GPa and 1.73 GPa for 500 h corrosion in liquid FLiNaK salt at 800 °C, 900 °C and 1000 °C, respectively. It seems that the corrosion process has no depravation on single fiber tensile strength. That is to say, the SiC fibers may have no contribute to the deterioration of mechanical performance of the composites. In terms of the crack growth, the gaps in corroded SiCf/SiC composites could deflect the cracks during fracture. The effect of corrosion gaps on crack growths during fracture are illustrated in Fig. 6. Although the fibers in the composites could improve the toughness of matrix, the microscale area in matrix around fiber bundles and between the fiber fabrics is still brittle[32], which makes the cracks grow directly through the microscale matrix and gives birth to brittle fracture surface in these microscale matrix (pointed by white arrow in Fig. 5a). After corrosion, the appearance of gaps resulted from the preferential corrosion of interlayer boundaries could deteriorate the bonding between deposited SiC matrix layers and damage the matrix integrity, which leads to the loss of matrix load capacity, layer exfoliation and corrosion damage caused by reaction. Under mechanical load, the short gaps emerged after corrosion at 800 °C could deflect cracks in matrix layer and give rise to the step-like matrix fracture surface (Fig. 6d). The thorough gaps achieved after corrosion at 900 °C aggravate the bonding force between the deposited matrix layers and lead to long length crack deflection, which results in pull-out of fiber bundle (Fig. 6f). At higher temperature (1000 °C), fibers can only 12
toughen the matrix partly because of the damage and exfoliation of matrix shown in Fig. 2d. The residual thin matrix layer around fibers bears little load and breaks as soon as the fibers fail, which results in the fiber bundle failure with short pull-out length (Fig. 6h). When the cracks propagate through the fibers and residual matrix layer, there are few matrix for further crack propagation and energy consumption, and the composites would fail at lower load. 5. Conclusions The microstructure evolution and high temperature mechanical properties of SiCf/SiC composites in LiFNaF-KF eutectic molten salt at various temperature for 500 h were studied. According to the experimental results, the following conclusions could be drawn: 1. SiC matrix interlayer boundaries with higher oxygen content are weak areas of the composites in terms of the corrosion resistance to fluoride molten salt. They are more liable to be corroded by the fluoride molten salt in comparison with the near-stoichiometric SiC layers. Different from the corrosion of CVD SiC, nonuniform corrosion of CVI SiCf/SiC composites in fluoride salt is observed. 2. Higher corrosion temperature leads to more severe damage of matrix. As the temperature increases, the composites suffer damage from part to entire corrosion of interlayer boundaries, and then matrix corrosion and layer exfoliation. Mass loss ratio induced by the molten salt corrosion shows an obvious increase with the increasing of corrosion temperature. XRD results suggest that the corrosion of SiC in fluoride salt is selective corrosion of Si, leading to the formation of graphite, which may be attributed to NO3- impurity in molten salt. 3. Because of the empty d orbitals in outermost electron shell of Si atom in SiC, the F ion, a kind of strong nucleophilic reagent, could provide common electrons to occupy the empty d orbitals to form Si-F coordinate bond. Si-C bond is replaced by Si-F bond, and hybridization form of Si shifts from sp3 to sp3d2. The Si bonds with F in salt to form [SiF6]2-, which dissolves into FLiNaK salt. The remained C4- would be oxidized by 13
impurities such as NO3- and then graphitized. 4. With the increase of corrosion temperature, the corroded samples suffer a severer deterioration in higher temperature flexural strength. The load capacity reduction of matrix and the oxidation of PyC interface during test may contribute to the deteriorated mechanical properties. SiCf/SiC composites after corrosion at 800 °C for 500 h retain about 51.6 % bending strength compared to the noncorroded samples. Acknowledgement The authors gratefully acknowledge the financial support from the Strategic Priority Research Programme of the Chinese Academy of Sciences (XDA02040203), the National Nature Science Foundation of China (No. 51502323) and the Natural Science Foundation of Shanghai, China (No. 14ZR1445800).
14
Reference [1] M. Gomina, P. Fourvel, M.-H. Rouillon, High temperature mechanical behaviour of an uncoated SiC-SiC composite material, Journal of Materials Science, 26 (1991) 1891-1898. [2] T.K. Jacobsen, P. Brøndsted, Mechanical properties of two plain‐ woven chemical vapor infiltrated silicon carbide‐ matrix composites, Journal of the American Ceramic Society, 84 (2001) 1043-1051. [3] K.Y. Donaldson, B.D. Trandell, Y. Lu, D. Hasselman, N. Maurer, Effect of delamination on the transverse thermal conductivity of a SiC‐ fiber‐ reinforced SiC‐ matrix composite, Journal of the American Ceramic Society, 81 (1998) 1583-1588. [4] W. Kowbel, C. Bruce, K. Tsou, K. Patel, J. Withers, G. Youngblood, High thermal conductivity SiC/SiC composites for fusion applications, Journal of nuclear materials, 283 (2000) 570-573. [5] N. Taylor, I. Cook, R. Forrest, C. Forty, W. Han, D. Ward, The safety, environmental and economic implications of the use of silicon carbide in fusion power plant blankets, Fusion Engineering and Design, 58 (2001) 991-995. [6] Y. Katoh, K. Ozawa, T. Hinoki, Y. Choi, L.L. Snead, A. Hasegawa, Mechanical properties of advanced SiC fiber composites irradiated at very high temperatures, Journal of Nuclear Materials, 417 (2011) 416-420. [7] Y. Katoh, T. Nozawa, L.L. Snead, K. Ozawa, H. Tanigawa, Stability of SiC and its composites at high neutron fluence, Journal of Nuclear Materials, 417 (2011) 400-405. [8] D.F. Williams, L.M. Toth, K.T. Clarno, Assessment of candidate molten salt coolants for the advanced high temperature reactor (AHTR), Oak Ridge National Lab., 2006 (Report ORNL/TM-2006/12). [9] M.B. Seregin, A.P. Parshin, A.Y. Kuznetsov, L.I. Ponomarev, S.A. Mel’nikov, A.A. Mikhalichenko, A.A. Rzheutskii, R.N. Manuilov, Solubility of UF4, ThF4, and CeF3 in a LiF-NaF-KF melt, Radiochemistry, 53 (2011) 491-493. 15
[10] D.F. Wilson, Chemical compatibility issues associated with use of SiC/SiC in advanced reactor concepts, Oak Bridge National Lab., 2012 (Report ORNL/TM-2012/546). [11] Y. Katoh, D. F. Wilson, C.W. Forsberg, Assessment of silicon carbide composites for advanced salt-cooled reactors, Oak Ridge National Lab., 2007 (Report ORNL/TM-2007/168). [12] X. Yang, D. Zhang, M. Liu, S. Feng, W. Xue, H. Liu, G. Yu, X. Zhou, H. Xia, P. Huai, Z. Li, Y. Lu, H. Zhou, S. Dong, Corrosion of SiC induced by Hastelloy N alloy and its corrosion products in LiF–NaF–KF molten salt, Corrosion Science, 109 (2016) 62-67. [13] W. Xue, X. Yang, J. Qiu, H. Liu, B. Zhao, H. Xia, X. Zhou, P. Huai, H. Liu, J. Wang, Effects of Cr 3+ on the corrosion of SiC in LiF–NaF–KF molten salt, Corrosion Science, 114 (2017) 96-101. [14] J. Schmidt, M. Scheiffele, M. Crippa, P.F. Peterson, E. Urquiza, K. Sridharan, L.C. Olson, M.H. Anderson, T.R. Allen, Y. Chen, Design, Fabrication, and Testing of Ceramic Plate‐ Type Heat Exchangers with Integrated Flow Channel Design, International Journal of Applied Ceramic Technology, 8 (2011) 1073-1086. [15] X. Yang, M. Liu, Y. Gao, D. Zhang, S. Feng, H. Liu, G. Yu, G. Wu, M. Wang, X. Zhou, H. Xia, P. Huai, T.K. Sham, J. Wang, J. Guo, Effect of oxygen on the corrosion of SiC in LiF–NaF–KF molten salt, Corrosion Science, 103 (2016) 165-172. [16] J.-S. Park, A. Kohyama, T. Hinoki, K. Shimoda, Y.-H. Park, Efforts on large scale production of NITESiC/SiC composites, Journal of Nuclear Materials, 367 (2007) 719-724. [17] Y. Katoh, M. Kotani, H. Kishimoto, W. Yang, A. Kohyama, Properties and radiation effects in hightemperature pyrolyzed PIP-SiC/SiC, Journal of Nuclear Materials, 289 (2001) 42-47. [18] S.P. Lee, J.S. Park, Y. Katoh, A. Kohyama, D.H. Kim, J.K. Lee, H.K. Yoon, Process, microstructure and flexural properties of reaction sintered Tyranno SA/SiC composites, Journal of Nuclear Materials, 307 (2002) 1191-1195. 16
[19] W. Krenkel, Ceramic Matrix Composites: Fiber Reinforced Ceramics and their Applications, John Wiley & Sons, Weinheim, Germany, 2008. [20] Y. Katoh, A. Kohyama, T. Hinoki, L.L. Snead, Progress in SiC-based ceramic composites for fusion applications, Fusion Science and Technology, 44 (2003) 155-162. [21] D. Buckthorpe, Material challenges for the next generation of fission reactor systems, in: J. LecomteBeckers, Q. Contrepois, T. Beck, B. Kuhn (Eds.) The 9th Liege Conference: Materials for Advanced Power Engineering, Forschungszentrum Juelich GmbH Liege, Belgium, 2010. [22] L.C. Olson, J.W. Ambrosek, K. Sridharan, M.H. Anderson, T.R. Allen, Materials corrosion in molten LiF–NaF–KF salt, Journal of Fluorine Chemistry, 130 (2009) 67-73. [23] L.C. Olson, J.W. Ambrosek, K. Sridharan, M.H. Anderson, T.R. Allen, M.L. Corradini, Evaluation of material corrosion in molten fluoride salt, in:
Presentation on AIChE conference, San-Francisco, 2006.
[24] T. Taguchi, T. Nozawa, N. Igawa, Y. Katoh, S. Jitsukawa, A. Kohyama, T. Hinoki, L.L. Snead, Fabrication of advanced SiC fiber/F-CVI SiC matrix composites with SiC/C multi-layer interphase, Journal of Nuclear Materials, 329 (2004) 572-576. [25] Y.P. Zaykov, A.V. Isakov, I.D. Zakiryanova, O.G. Reznitskikh, O.V. Chemezov, A.A. Redkin, Interaction between SiO2 and a KF–KCl–K2SiF6 Melt, The Journal of Physical Chemistry B, 118 (2014) 1584-1588. [26] Y. Wang, Z.-F. Tang, Y. Fu, S.-R. Huang, S.-F. Zhao, P. Zhang, L.-D. Xie, X.-G. Wang, G.-J. Zhang, Corrosion behavior of ZrC–SiC composite ceramics in LiF–NaF–KF molten salt at high temperatures, Ceramics International, 41 (2015) 12996-13005. [27] A. Ouerghi, R. Belkhou, M. Marangolo, M.G. Silly, S. El Moussaoui, M. Eddrief, L. Largeau, M. Portail, F. Sirotti, Structural coherency of epitaxial graphene on 3C–SiC(111) epilayers on Si(111), Applied Physics Letters, 97 (2010) 161905. 17
[28] M. Suemitsu, Y. Miyamoto, H. Handa, A. Konno, Graphene formation on a 3C-SiC(111) thin film grown on Si(110) substrate, e-Journal of Surface Science and Nanotechnology, 7 (2009) 311-313. [29] A.J. Van Bommel, J.E. Crombeen, A. Van Tooren, LEED and Auger electron observations of the SiC(0001) surface, Surface Science, 48 (1975) 463-472. [30] L.M. Zhang, X.H. Huang, X.L. Song, Fundamentals of Materials Science, Wuhan University of Technology Press, Wuhan, 2004. [31] G. Zhenduo, Z. Zhongtai, J. Jinsheng, Physical Properties of Inorganic Materials, Tsinghua University Press, Beijing, 1992. [32] W. Yang, H. Araki, A. Kohyama, S. Thaveethavorn, H. Suzuki, T. Noda, Process and mechanical properties of in situ silicon carbide-nanowire-reinforced chemical vapor infiltrated silicon carbide/silicon carbide composite, Journal of the American Ceramic Society, 87 (2004) 1720-1725.
18
Fig. 1. Surface SEM images of SiCf/SiC composites (a) before corrosion and after corrosion in liquid FLiNaK salt at (b) 800 °C, (c) 900 °C and (d) 1000 °C for 500 h.
19
Fig. 2. Cross-section SEM images of SiCf/SiC composites (a) before corrosion and after corrosion in liquid FLiNaK salt at (b) 800°C, (c) 900 °C and (d) 1000 °C for 500 h. The insert of (a) is the EDS liner scan result of oxygen element cross the boundary between deposited SiC layers in SiC matrix.
20
Fig. 3. XRD patterns of SiCf/SiC composites (a) before and (b) after corrosion in liquid FLiNaK salt at varying temperatures for 500 h.
21
Fig. 4. Stress - strain curves of SiCf/SiC composites before and after corrosion. Samples corroded in FLiNaK salt at corrosion temperature and ruptured at flexural temperature are marked as “corrosion temperature @ flexural temperature”. The “Blank” is the representation of as-received samples.
22
Fig. 5. Fracture surface SEM images of (a) as-received SiCf/SiC composites ruptured at 800 °C and corroded SiCf/SiC composites in FLiNaK salt (corrosion temperature @ fracture temperature): (b) 800 °C @ 800 °C, (c) 900 °C @ 900 °C and (d) 1000 °C @ 1000 °C.
23
Fig. 6. Illustration of microstructure evolution and crack growth during rupture in SiCf/SiC composites before and after corrosion in FLiNaK salt (corrosion temperature @ fracture temperature): (a)(b) as-received samples, (c)(d) 800 °C @ 800 °C, (e)(f) 900 °C @ 900 °C and (g)(h) 1000 °C @ 1000 °C.
24
Table 1 Mass loss ratio of SiCf/SiC composites after immersion corrosion in liquid FLiNaK salt at various temperature for 500 h. Every bar in each experiment group was weighed 3 times independently to obtain the weight of each bar, then use the data of 5 bars in each experiment group to calculate the average mass loss ratio, mass loss rate and the corresponding standard deviation. Experimental Group
800 °C-500 h
900 °C-500 h
1000 °C-500 h
Mass loss ratio / %
5.914±0.578
15.144±0.642
31.062±0.707
Mass loss rate / μg·mm-2·h-1
0.211±0.021
0.537±0.023
1.107±0.024
25
Table 2 High temperature flexural strength of SiCf/SiC composites before and after corrosion in liquid FLiNaK salt at varying temperatures. Each experiment group tested 3 bars to calculate the average flexural strength and standard deviation. Flexural strength / MPa Flexural temperature / °C
Flexural strength retention ratio / % Before corrosion
After corrosion*
800
450.60 ± 90.49
232.67 ± 44.68
51.64
900
270.86 ± 26.82
176.77 ± 13.33
65.26
1000
238.89 ± 4.63
60.03 ± 11.81
25.13
* The corrosion temperature is as same as the flexural temperature.
26
Table 3 Porosity of SiCf/SiC composites before and after corrosion in liquid FLiNaK salt at varying temperature for 500 h. Each group tested 5 bars to calculate the average porosity and standard deviation Corrosion temperature / °C Sample
Porosity / %
As-received
5.01±0.13
800
900
1000
5.37±0.57
6.73±0.38
12.42±1.88
27
S0010-938X(16)30734-X http://dx.doi.org/doi:10.1016/j.corsci.2017.05.016 CS 7092
To appear in: Received date: Revised date: Accepted date:
10-9-2016 16-5-2017 23-5-2017
Please cite this article as: Hongda Wang, Qian Feng, Zhen Wang, Haijun Zhou, Yanmei Kan, Jianbao Hu, Shaoming Dong, Microstructure evolution and high-temperature mechanical properties of SiCf/SiC composites in liquid fluoride salt environment, Corrosion Sciencehttp://dx.doi.org/10.1016/j.corsci.2017.05.016 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Microstructure evolution and high-temperature mechanical properties of SiCf/SiC composites in liquid fluoride salt environment Hongda Wanga,b,c, Qian Fengd, Zhen Wanga,b,, Haijun Zhoua,b,*, Yanmei Kana,b, Jianbao Hua,b, Shaoming Donga,b,* a
State Key Laboratory of High Performance Ceramics and Superfine Microstructure, Shanghai Institute of
Ceramics, Chinese Academy of Sciences, Shanghai 200050, China. b
Structural Ceramics and Composites Engineering Research Center, Shanghai Institute of Ceramics, Chinese
Academy of Sciences, Shanghai 200050, China. c
University of Chinese Academy of Sciences, Beijing 100049, China.
d
Analysis and Testing Center, Donghua University, Shanghai 201600, China.
*Corresponding author: Haijun Zhou: Tel.: +86 21 69906076
Fax: +86 21 69906085
E-mail: [email protected] Address: Shanghai Institute of Ceramics, Chinese Academy of Sciences, No. 1295, Dingxi Road, Changning District, Shanghai 200050, P. R. China. Shaoming Dong: Tel.: +86 21 52414324
Fax: +86 21 69906085
E-mail: [email protected] Address: Shanghai Institute of Ceramics, Chinese Academy of Sciences, No. 1295, Dingxi Road, Changning District, Shanghai 200050, P. R. China.
1
Highlights:
Corrosion behavior of SiCf/SiC CMCs in fluoride salt was studied.
O-contained boundaries result in the non-uniform corrosion of SiC matrix.
The Si-F coordinate bond plays important roles in SiC corrosion.
Corrosion damage of matrix leads to degradation of mechanical properties.
Abstract High temperature mechanical properties and microstructure evolution of chemical-vapor-infiltrated SiC fiber reinforced SiC ceramic matrix composites (SiCf/SiC) in 46.5LiF-11.5NaF-42.0KF (mol. %, FLiNaK) eutectic salt were investigated. The results indicate that the corrosion of SiCf/SiC composites was accelerated with increase of corrosion temperature. Interlayer boundaries in SiC matrix with higher oxygen content were corroded preferentially. The F- formed Si-F coordinate bond with SiC and replaced carbon, followed by the corrosion of SiC matrix. While single SiC fiber tensile strength suffered no depravation after corrosion, corrosion and damage of SiC matrix led to deteriorated high temperature mechanical properties of SiCf/SiC composites.
Key words: A. Ceramic matrix composites; A. Molten salt; B. SEM; B. XRD; C. Exfoliation corrosion; C. High temperature corrosion
1. Introduction Silicon carbide (SiC) continuous fiber-reinforced SiC matrix composites (SiCf/SiC) have a multitude of excellent properties, including high temperature mechanical performance with better mechanical damage tolerance over inherently brittle monolithic SiC[1, 2], outstanding thermal conductivity, low activation to 2
neutron-induced radioactivity[3-7], which make the SiCf/SiC composites a promising candidate material for structural and functional components in nuclear reactors. Particularly, SiCf/SiC composites have nice compatibility with liquid fluoride salt (e.g. 46.5 mol % LiF- 11.5 mol % NaF-42.0 mol % KF eutectic molten salt, i.e. FLiNaK molten salt). This unique properties of SiCf/SiC composites make it achieve wide attention due to their potential application in heat-exchanger, control rod and other structure components in Molten Salt Reactor (MSR)[8-11]. However, SiC undergoes corrosion in liquid fluoride salt at high temperature because of the disequilibrium in thermodynamics between SiC and fluoride salt[11]. Obviously, the corrosion of SiC compromises the advantages of SiCf/SiC composites in nuclear application. The corrosion behavior of SiCf/SiC composites in liquid fluoride salt has been seldom investigated up to now, especially the evolution of microstructure and mechanical performance. It is indicated that the impurities in fluoride salt and the composition of SiC-based materials lead to the incompatibility of SiCf/SiC composites to fluoride salt. The foreign ions (i.e. Ni2+, Cr2+ and Cr3+), which could be introduced during production of fluoride salt or by corrosion of Hastelloy N alloy, can accelerate the corrosion of SiC[12, 13]. Although the reason of optionally corrosion of SiC is still unknown, qualitative explanation with the respect of the standard Gibbs free energy of fluoride formation per F2 molecule of the Si and SiC phases has been widely accepted[14]. The Si-O bond in O-contained SiC phase would be attacked by F- to form Si-F bond, indicating that the oxygen in SiC accelerates the corrosion of SiC in fluoride salt[15]. Therefore, high-purity SiC is needed in nuclear reactor application. There is a variety of technology for preparing SiCf/SiC composites, such as nano-infiltration and transient eutectic phase[16], polymer impregnation and pyrolysis[17], reaction sintering[18] and chemical vapor infiltration (CVI) process[19]. Among these methods, the CVI process could form highly crystalline, nearstoichiometric SiC matrix and minimize damage of fibers during material fabrication, which are beneficial to 3
produce the nuclear grade SiCf/SiC composites[20]. However, few research has focused on the corrosion behavior of CVI SiCf/SiC composites and mechanical performance in liquid fluoride salt environment. In this work, the static corrosion behavior of CVI SiCf/SiC composites in molten FLiNaK salt at various temperature was studied. The microstructure evolution during the corrosion process, and the high temperature mechanical properties were investigated. 2. Experimental The materials used to evaluate corrosion behavior are SiC matrix composites reinforced by two-dimensional woven fabrics of KD-II SiC fibers (National University of Defense Technology, Changsha, China). Using the Methyltrichlorosilane (MTS) as the raw material and H2 as the carrier gas, the SiCf/SiC composites were prepared via CVI processing. Prior to the CVI process, a pyrolytic carbon (PyC) monolayer of ~ 100 nm in thickness was deposited on the SiC fiber surface through chemical vapor deposition. The fiber volume fraction, porosity and density of the as-received composites were ~ 40%, ~ 5.01 % and 2.74 g∙cm-3, respectively. The monolithic composites were machined into 35 × 4 × 2 mm3 bar for the corrosion experiment. Each test group contained 5 bars. According to the design goals of Generation IV reactors, the outlet temperature of MSR would be 700 °C ~ 800 °C[21], and the peak temperature of Advanced High Temperature Reactor would be up to 1000 °C[11]. To evaluate the performance of SiCf/SiC composites in the similar environment in reactor, 800 °C, 900 °C and 1000 °C were chosen as the corrosion temperature, and the corrosion time was 500 h[22]. The SiCf/SiC composites samples were transferred into a glove box under the protection of high purity argon gas. The atmosphere in the glove box was controlled at O2 < 0.1 ppm and moisture < 0.1 ppm all along. Before corrosion, the graphite crucible and SiCf/SiC samples were heat-treated in furnace in the glove box at 700 °C for 10 h to eliminate possible adsorption of air and other gases. Subsequently, SiCf/SiC samples were fixed on a graphite 4
support and dipped into ~ 80 g 46.5LiF-11.5NaF-42.0KF (mol. %, FLiNaK) fluoride eutectic molten salt (Shanghai Institute of Organic Chemistry, Chinese Academy of Sciences, Shanghai) in the graphite crucibles. The same SiC fibers used in SiCf/SiC composites were also introduced into the crucibles at the same time to study the effect of corrosion on mechanical properties of SiC fibers. And then, the crucibles were heated to 800 °C, 900 °C or 1000 °C for 500 h, respectively. After the high temperature corrosion process in fluoride eutectic molten salt, the SiCf/SiC composite samples were picked out from the crucible and immersed into 1 mol∙L-1 Al(NO3)3 solution to remove the fluoride salt solid adhered on the exterior and interior surfaces of the sample[23]. Finally, the samples were washed with deionized water using an ultrasonic cleaner, and then dried at 120 °C in a vacuum drying chamber. The samples were weighed by precision balance (BT-25-S, 0.01 mg, Sartorius) before and after the corrosion test, and the mass loss ratio was calculated to characterize the corrosion resistance of the composites. Every bar in each test group was measured 3 times independently to obtain the mass loss ratio and mass loss rate of each bar. Then data of 5 bars in each experiment group was used to calculate the average mass loss ratio, mass loss rate and the corresponding standard deviation. The mass loss ratio (ml) could be calculated by the formula below:
ml
m0 m 100% m0
(1)
where m0 and m are the mass of sample before and after corrosion, respectively. The apparent mass loss rate (v) was introduced to semi-quantitatively describe the corrosion rate:
m0 m m0 m S t 2( LH LW HW ) t
(2)
where S is surface area of sample, t is corrosion time, and L, W, H are the length, width, height of sample, respectively. Phase compositions of the samples were analyzed by X-ray diffraction (XRD, D8 Advance, 5
Bruker AXS Co. Ltd., GER) at room temperature, using Cu Kα radiation (λ = 0.154 nm) 2θ over the angles range of 10 ° ~ 80 °. High temperature flexural strength of the samples after corrosion was measured through a three-point-bending test (Instron-5500R, Instron Corp., Canton, USA) with a span of 30 mm and a loading rate of 0.5 mm/min in air atmosphere. For a comparison, the high temperature flexural strength of samples without corrosion was also measured. 3 bars were tested for each group. Polished cross-section and surface morphologies of the samples before and after corrosion at varying temperatures were characterized by scanning electron microscope (SEM, Quanta-250, FEI, USA) along with energy dispersive spectrometer (EDS, Oxford, UK). Additionally, the single fiber tensile strength of SiC fibers before and after corrosion was tested by single fiber electronic tensile strength tester (YG-001A, Hongda Fangyuan Electric Co. Ltd., China), using a span distance of 20 mm and stretching rate of 15 mm·min-1. 3. Results 3.1. Morphologies of SiCf/SiC composites The surface morphologies of as-received and corroded SiCf/SiC composites samples are given in Fig. 1. In comparison with the planar surface of untreated samples, rougher surface is observed after corrosion. After exposed in liquid FLiNaK salt at 800 °C for 500 h, matrix corrosion and exfoliation occurs. At higher temperature of 900 °C and 1000 °C, more intensive destruction, including extensive matrix layer exfoliation around fibers and fiber damage, was observed clearly. Polished cross-section microstructures of as-received and corroded samples at varying temperatures are shown in Fig. 2. Matrix exhibits typical feature of layer structure, which is resulted from repetition of CVI process. Also, residual pores can be observed in fiber bundle (Fig. 2a). Line scan result given in white box in Fig. 2a shows that interlayer boundaries contain slightly more oxygen than matrix. After corrosion in liquid fluoride salt at varying temperatures, the interlayer boundaries between deposited matrix layers are selectively 6
corroded, which leads to gaps formed between deposited layers. When corroded at 800 °C for 500 h, parts of interlayer boundaries and matrix around residual pores are eroded (pointed by arrows in Fig. 2b). After corrosion at 900 °C, interlayer boundaries between deposited matrix layers are almost completely corroded. Layer exfoliation is observed around the interior surface of residual pores. Corrosion at 1000 °C causes intensive damage of matrix, especially the serious layer exfoliation of deposited matrix layers. Mechanical damage of matrix around pores and corroded boundaries during polishing process is also found because of loose structure of matrix caused by corrosion. Owing to the formation of gaps in matrix resulted from the selective corrosion of interlayer boundaries, the molten salt would penetrate into the composites, leading to extension of corrosion in both exterior and interior of the sample. The microstructure evolution before and after corrosion indicates a non-uniform corrosion, which was different from the uniform corrosion of CVD SiC in fluoride salt reported in references [12, 15]. 3.2. Mass change Considering the non-uniform corrosion process (especially the optional corrosion of interlayer boundaries) caused by the anisotropic microstructure and heterogeneous composition of the composites, the normal expression on corrosion by using a mass loss per unit area or a thickness may not be appropriate. Therefore, the mass loss ratio is used to describe the corrosion damage degree of samples. At the same time, the apparent mass loss rate is employed to characterize the corrosion rate semi-quantitatively because of the change of corrosion surface area during corrosion. Table 1 shows the mass loss ratio and apparent mass loss rate of samples corroded at 800 °C, 900 °C and 1000 °C. As the corrosion temperature rises from 800 °C to 1000 °C, the mass loss ratio increases from 5.914 wt. % to 31.062 wt. %, while the apparent mass loss rate raises from 0.211 μg·mm-2·h-1 to 1.107 μg·mm-2·h-1, which suggests that higher corrosion temperature would aggravate the damage of SiCf/SiC composites in liquid fluoride salt. 7
3.3. X-ray diffraction pattern Fig. 3 shows the XRD patterns of the SiCf/SiC composites before and after corrosion at selected temperatures for 500 h. Different from the XRD pattern of as-received SiCf/SiC composites, diffraction peaks corresponding to Graphite-2H are detected in XRD patterns of the corroded samples. Peak intensity of the graphite phase increases with the corrosion temperature. Graphite-2H is believed to be the corrosion product of β-SiC in the molten salt. 3.4. High temperature mechanical property High temperature mechanical properties of SiCf/SiC composites under corrosive environment are important and fundamental for the application in salt-cooled reactors. Three-point bending test at corresponding corrosion temperature was employed to characterize the effect of corrosion in liquid fluoride salt on mechanical performance of SiCf/SiC composites. The flexural strength of samples before and after corrosion at varying temperatures are listed in Table 2. Compared with the flexural strength of as-received samples tested at corresponding corrosion temperatures, the flexural strength of samples after corrosion exhibits significant decrement. Flexural strength retention ratio of the samples after corrosion at 800 °C, 900 °C and 1000 °C compared to that of the as-received samples at corresponding temperature were 51.64 %, 65.26 % and 25.13, respectively. Both the flexural strength before and after corrosion decreases with the increase of the testing temperature, and the corroded samples suffer larger drop. Table 3 gives the porosity of each group samples before and after corrosion. Corrosion process leads to higher porosity, and larger sized pores are obtained at higher corrosion temperature. The high temperature stress - strain curves of the samples before and after corrosion are presented in Fig. 4. All curves contain the extended regions after the initial failure, which indicate the ductile rupture behavior for all the samples before and after corrosion. 8
Fig. 5 shows fracture surface morphologies of the SiCf/SiC composites. Long pull-out fibers are observed on the fracture surface of blank samples and corroded samples at 800 °C, while long pull-out fiber bundle instead of common pull-out fibers occurs at 900 °C and fiber bundle breaks with short finite length at 1000 °C. This is typical of crack arresting, deflecting and branching behavior which leads to the ductile rupture mode of the composites. Interestingly, compared with the brittle fracture matrix around the fiber bundle in blank sample, the corroded sample at 800 °C shows step-like morphology on fracture surface, which is similar to the fracture morphology of (SiC/PyC) multilayer[24]. 4. Discussion 4.1. Corrosion behavior The corrosion of SiC in fluoride salt is very complex, and the mechanism is still unknown. It is well known that the oxygen in SiC as well as the impurities in fluoride salt shows great influence on the corrosion behavior [12, 13, 15]. Compared with the CVD SiC, the SiCf/SiC composites with different microstructure may show distinctive corrosion behavior. As shown in Fig. 2, the appearance of gaps resulted from the preferential corrosion of interlayer boundaries indicates that the corrosion of SiCf/SiC composites would be non-uniform corrosion. EDS linear scan results shown in Fig. 2a demonstrates that oxygen content in interlayer boundaries is slightly higher than that in matrix. This may be associated with the repetition of densification process. To decrease the porosity of SiCf/SiC composites, component surface usually needs to be grinded to reopen the pores that sealed off during deposition, which guarantees the reactants to penetrate into the component during the subsequent infiltration process. However, oxygen absorption and slow oxidation would occur on the surface of deposited SiC layer when reopening the pores. The O-rich surface can be covered by subsequent deposited SiC layer, forming interlayer boundaries with high oxygen content between deposited SiC layers. Higher oxygen content leads to 9
worse corrosion resistance of interlayer boundaries than near-stoichiometric SiC[15]. When dipped into FLiNaK salt, Si-O bond in these interlayers would be attacked by F- to form more stable Si-F bond[25]. SiF4, which is believed to be the corrosion product, is considered to dissolve in FLiNaK salt and form SiF62-. As a result, the interlayer boundaries would be corroded into gaps, which destructs the bonding force between deposited layers and causes the layer exfoliation (Fig. 2c). At the same time, liquid salt infiltrates into the matrix through the corrosion gaps and further damages the matrix inside the composites (Fig. 2d). The XRD patterns shown in Fig. 3 also suggests that the corrosion of SiC in fluoride salt would be the loss of Si in SiC, with graphite as the corrosion product. The F ion has large electronegativity and is a kind of strong nucleophilic reagent. The outermost electron shell of Si atom has 4 electrons, which would form covalent bond with valence electrons of C atom in SiC through sp3 hybrid. However, there are still empty electronic orbits, d orbitals, in outermost electron shell of Si. During the corrosion, F- would provide common electrons to occupy the empty d orbitals of Si to form Si-F coordinate bond. This Si-F bond could change the symmetrical structure of [SiC4] tetrahedron and is conducive to further replacement of C by F. As a result, SiC bond is replaced by Si-F bond, and hybridization form of Si shifts from sp3 to sp3d2. The Si bonds with F in salt to form [SiF6]2-, which dissolved into FLiNaK salt[15]. The [SiF6]2- could decompose into SiF4 gas and escape out from liquid salt at high temperature, which contributes to the further corrosion and dissolution of SiC. At the same time, the C4- in SiC would be oxidized into C by impurities, such as NO3-[26]. And then, newly formed C may crystallizes into graphite, which might be similar to the graphitization process on the surface of SiC[27, 28]. After the dissolution of Si in form of [SiF6]2-, two or three successive carbon layers collapse into one layer to form graphite[29]. From Fig. 1 and Fig. 2, heavier corrosion damage of SiCf/SiC composites is observed at higher temperature, indicating that higher temperature would accelerate the corrosion of SiCf/SiC composites. This is coincident 10
with the mass loss ratio in Table 1, the XRD patterns in Fig. 3 obtained at selected temperatures and the porosity of samples shown in Table 3. According to the Arrhenius equation[30], higher temperature supplies enough energy for overcoming the reaction barrier, known as the activation energy (E a), and more effective collision would happen during reaction. The corrosion reaction would be accelerated at higher temperature. Consequently, this causes worse damage and larger mass loss ratio at higher corrosion temperature. 4.2 Mechanical properties Considering the microstructure evolution of samples corroded at varying temperature shown in Fig. 1 and Fig. 2, the serious decrement of high temperature flexural strength with increase of the corrosion temperature would be attributed to the corrosion damage caused by fluoride salt. Initially, the selective corrosion of O-contained interlayer boundaries destroys the integrity of matrix and lowers the capacity of matrix. As shown in Fig. 4, the slop of ascending portion in stress - stain curves of corroded samples decreases with increasing corrosion temperature, and is also smaller than that of as-received samples obtained at corresponding temperature. This suggests that the elastic modulus of samples degrades in corrosion process. Gaps resulted from the selective corrosion of interlayer boundaries should be responsible for this. As shown in Table 1, the appearance of gaps in matrix after corrosion increases the porosity in the composites. The relationship between elastic modulus and porosity could be expressed as the empirical equation below[31]:
E E0 (1 1.9P 0.9P2 )
(3)
where P is the porosity of material, E is the elastic modulus of material when the porosity is P, and E0 is the elastic modulus of fully dense material. After corrosion, higher the corrosion temperature is, larger porosity the sample would obtain, which lowers the modulus significantly. As a result, matrix load capacity would decrease and mechanical performance is deteriorated. 11
Secondly, the oxidation of PyC interface in SiCf/SiC composites during test may partly account for the degradation of mechanical properties. The high temperature flexural strength test was conducted in air atmosphere. The PyC interface would react with O2 during heating, and the oxidation rate of PyC increases with increased temperature. The damage of PyC interface would hinder the load transfer between SiC matrix and SiC fibers. In consequence, the flexural strength would decrease at higher temperature. Additionally, corrosion experiments of the same SiC fibers used in composites were conducted. Single fiber tensile strength of the SiC fibers is 1.72 GPa before corrosion and is 1.65 GPa, 1.73 GPa and 1.73 GPa for 500 h corrosion in liquid FLiNaK salt at 800 °C, 900 °C and 1000 °C, respectively. It seems that the corrosion process has no depravation on single fiber tensile strength. That is to say, the SiC fibers may have no contribute to the deterioration of mechanical performance of the composites. In terms of the crack growth, the gaps in corroded SiCf/SiC composites could deflect the cracks during fracture. The effect of corrosion gaps on crack growths during fracture are illustrated in Fig. 6. Although the fibers in the composites could improve the toughness of matrix, the microscale area in matrix around fiber bundles and between the fiber fabrics is still brittle[32], which makes the cracks grow directly through the microscale matrix and gives birth to brittle fracture surface in these microscale matrix (pointed by white arrow in Fig. 5a). After corrosion, the appearance of gaps resulted from the preferential corrosion of interlayer boundaries could deteriorate the bonding between deposited SiC matrix layers and damage the matrix integrity, which leads to the loss of matrix load capacity, layer exfoliation and corrosion damage caused by reaction. Under mechanical load, the short gaps emerged after corrosion at 800 °C could deflect cracks in matrix layer and give rise to the step-like matrix fracture surface (Fig. 6d). The thorough gaps achieved after corrosion at 900 °C aggravate the bonding force between the deposited matrix layers and lead to long length crack deflection, which results in pull-out of fiber bundle (Fig. 6f). At higher temperature (1000 °C), fibers can only 12
toughen the matrix partly because of the damage and exfoliation of matrix shown in Fig. 2d. The residual thin matrix layer around fibers bears little load and breaks as soon as the fibers fail, which results in the fiber bundle failure with short pull-out length (Fig. 6h). When the cracks propagate through the fibers and residual matrix layer, there are few matrix for further crack propagation and energy consumption, and the composites would fail at lower load. 5. Conclusions The microstructure evolution and high temperature mechanical properties of SiCf/SiC composites in LiFNaF-KF eutectic molten salt at various temperature for 500 h were studied. According to the experimental results, the following conclusions could be drawn: 1. SiC matrix interlayer boundaries with higher oxygen content are weak areas of the composites in terms of the corrosion resistance to fluoride molten salt. They are more liable to be corroded by the fluoride molten salt in comparison with the near-stoichiometric SiC layers. Different from the corrosion of CVD SiC, nonuniform corrosion of CVI SiCf/SiC composites in fluoride salt is observed. 2. Higher corrosion temperature leads to more severe damage of matrix. As the temperature increases, the composites suffer damage from part to entire corrosion of interlayer boundaries, and then matrix corrosion and layer exfoliation. Mass loss ratio induced by the molten salt corrosion shows an obvious increase with the increasing of corrosion temperature. XRD results suggest that the corrosion of SiC in fluoride salt is selective corrosion of Si, leading to the formation of graphite, which may be attributed to NO3- impurity in molten salt. 3. Because of the empty d orbitals in outermost electron shell of Si atom in SiC, the F ion, a kind of strong nucleophilic reagent, could provide common electrons to occupy the empty d orbitals to form Si-F coordinate bond. Si-C bond is replaced by Si-F bond, and hybridization form of Si shifts from sp3 to sp3d2. The Si bonds with F in salt to form [SiF6]2-, which dissolves into FLiNaK salt. The remained C4- would be oxidized by 13
impurities such as NO3- and then graphitized. 4. With the increase of corrosion temperature, the corroded samples suffer a severer deterioration in higher temperature flexural strength. The load capacity reduction of matrix and the oxidation of PyC interface during test may contribute to the deteriorated mechanical properties. SiCf/SiC composites after corrosion at 800 °C for 500 h retain about 51.6 % bending strength compared to the noncorroded samples. Acknowledgement The authors gratefully acknowledge the financial support from the Strategic Priority Research Programme of the Chinese Academy of Sciences (XDA02040203), the National Nature Science Foundation of China (No. 51502323) and the Natural Science Foundation of Shanghai, China (No. 14ZR1445800).
14
Reference [1] M. Gomina, P. Fourvel, M.-H. Rouillon, High temperature mechanical behaviour of an uncoated SiC-SiC composite material, Journal of Materials Science, 26 (1991) 1891-1898. [2] T.K. Jacobsen, P. Brøndsted, Mechanical properties of two plain‐ woven chemical vapor infiltrated silicon carbide‐ matrix composites, Journal of the American Ceramic Society, 84 (2001) 1043-1051. [3] K.Y. Donaldson, B.D. Trandell, Y. Lu, D. Hasselman, N. Maurer, Effect of delamination on the transverse thermal conductivity of a SiC‐ fiber‐ reinforced SiC‐ matrix composite, Journal of the American Ceramic Society, 81 (1998) 1583-1588. [4] W. Kowbel, C. Bruce, K. Tsou, K. Patel, J. Withers, G. Youngblood, High thermal conductivity SiC/SiC composites for fusion applications, Journal of nuclear materials, 283 (2000) 570-573. [5] N. Taylor, I. Cook, R. Forrest, C. Forty, W. Han, D. Ward, The safety, environmental and economic implications of the use of silicon carbide in fusion power plant blankets, Fusion Engineering and Design, 58 (2001) 991-995. [6] Y. Katoh, K. Ozawa, T. Hinoki, Y. Choi, L.L. Snead, A. Hasegawa, Mechanical properties of advanced SiC fiber composites irradiated at very high temperatures, Journal of Nuclear Materials, 417 (2011) 416-420. [7] Y. Katoh, T. Nozawa, L.L. Snead, K. Ozawa, H. Tanigawa, Stability of SiC and its composites at high neutron fluence, Journal of Nuclear Materials, 417 (2011) 400-405. [8] D.F. Williams, L.M. Toth, K.T. Clarno, Assessment of candidate molten salt coolants for the advanced high temperature reactor (AHTR), Oak Ridge National Lab., 2006 (Report ORNL/TM-2006/12). [9] M.B. Seregin, A.P. Parshin, A.Y. Kuznetsov, L.I. Ponomarev, S.A. Mel’nikov, A.A. Mikhalichenko, A.A. Rzheutskii, R.N. Manuilov, Solubility of UF4, ThF4, and CeF3 in a LiF-NaF-KF melt, Radiochemistry, 53 (2011) 491-493. 15
[10] D.F. Wilson, Chemical compatibility issues associated with use of SiC/SiC in advanced reactor concepts, Oak Bridge National Lab., 2012 (Report ORNL/TM-2012/546). [11] Y. Katoh, D. F. Wilson, C.W. Forsberg, Assessment of silicon carbide composites for advanced salt-cooled reactors, Oak Ridge National Lab., 2007 (Report ORNL/TM-2007/168). [12] X. Yang, D. Zhang, M. Liu, S. Feng, W. Xue, H. Liu, G. Yu, X. Zhou, H. Xia, P. Huai, Z. Li, Y. Lu, H. Zhou, S. Dong, Corrosion of SiC induced by Hastelloy N alloy and its corrosion products in LiF–NaF–KF molten salt, Corrosion Science, 109 (2016) 62-67. [13] W. Xue, X. Yang, J. Qiu, H. Liu, B. Zhao, H. Xia, X. Zhou, P. Huai, H. Liu, J. Wang, Effects of Cr 3+ on the corrosion of SiC in LiF–NaF–KF molten salt, Corrosion Science, 114 (2017) 96-101. [14] J. Schmidt, M. Scheiffele, M. Crippa, P.F. Peterson, E. Urquiza, K. Sridharan, L.C. Olson, M.H. Anderson, T.R. Allen, Y. Chen, Design, Fabrication, and Testing of Ceramic Plate‐ Type Heat Exchangers with Integrated Flow Channel Design, International Journal of Applied Ceramic Technology, 8 (2011) 1073-1086. [15] X. Yang, M. Liu, Y. Gao, D. Zhang, S. Feng, H. Liu, G. Yu, G. Wu, M. Wang, X. Zhou, H. Xia, P. Huai, T.K. Sham, J. Wang, J. Guo, Effect of oxygen on the corrosion of SiC in LiF–NaF–KF molten salt, Corrosion Science, 103 (2016) 165-172. [16] J.-S. Park, A. Kohyama, T. Hinoki, K. Shimoda, Y.-H. Park, Efforts on large scale production of NITESiC/SiC composites, Journal of Nuclear Materials, 367 (2007) 719-724. [17] Y. Katoh, M. Kotani, H. Kishimoto, W. Yang, A. Kohyama, Properties and radiation effects in hightemperature pyrolyzed PIP-SiC/SiC, Journal of Nuclear Materials, 289 (2001) 42-47. [18] S.P. Lee, J.S. Park, Y. Katoh, A. Kohyama, D.H. Kim, J.K. Lee, H.K. Yoon, Process, microstructure and flexural properties of reaction sintered Tyranno SA/SiC composites, Journal of Nuclear Materials, 307 (2002) 1191-1195. 16
[19] W. Krenkel, Ceramic Matrix Composites: Fiber Reinforced Ceramics and their Applications, John Wiley & Sons, Weinheim, Germany, 2008. [20] Y. Katoh, A. Kohyama, T. Hinoki, L.L. Snead, Progress in SiC-based ceramic composites for fusion applications, Fusion Science and Technology, 44 (2003) 155-162. [21] D. Buckthorpe, Material challenges for the next generation of fission reactor systems, in: J. LecomteBeckers, Q. Contrepois, T. Beck, B. Kuhn (Eds.) The 9th Liege Conference: Materials for Advanced Power Engineering, Forschungszentrum Juelich GmbH Liege, Belgium, 2010. [22] L.C. Olson, J.W. Ambrosek, K. Sridharan, M.H. Anderson, T.R. Allen, Materials corrosion in molten LiF–NaF–KF salt, Journal of Fluorine Chemistry, 130 (2009) 67-73. [23] L.C. Olson, J.W. Ambrosek, K. Sridharan, M.H. Anderson, T.R. Allen, M.L. Corradini, Evaluation of material corrosion in molten fluoride salt, in:
Presentation on AIChE conference, San-Francisco, 2006.
[24] T. Taguchi, T. Nozawa, N. Igawa, Y. Katoh, S. Jitsukawa, A. Kohyama, T. Hinoki, L.L. Snead, Fabrication of advanced SiC fiber/F-CVI SiC matrix composites with SiC/C multi-layer interphase, Journal of Nuclear Materials, 329 (2004) 572-576. [25] Y.P. Zaykov, A.V. Isakov, I.D. Zakiryanova, O.G. Reznitskikh, O.V. Chemezov, A.A. Redkin, Interaction between SiO2 and a KF–KCl–K2SiF6 Melt, The Journal of Physical Chemistry B, 118 (2014) 1584-1588. [26] Y. Wang, Z.-F. Tang, Y. Fu, S.-R. Huang, S.-F. Zhao, P. Zhang, L.-D. Xie, X.-G. Wang, G.-J. Zhang, Corrosion behavior of ZrC–SiC composite ceramics in LiF–NaF–KF molten salt at high temperatures, Ceramics International, 41 (2015) 12996-13005. [27] A. Ouerghi, R. Belkhou, M. Marangolo, M.G. Silly, S. El Moussaoui, M. Eddrief, L. Largeau, M. Portail, F. Sirotti, Structural coherency of epitaxial graphene on 3C–SiC(111) epilayers on Si(111), Applied Physics Letters, 97 (2010) 161905. 17
[28] M. Suemitsu, Y. Miyamoto, H. Handa, A. Konno, Graphene formation on a 3C-SiC(111) thin film grown on Si(110) substrate, e-Journal of Surface Science and Nanotechnology, 7 (2009) 311-313. [29] A.J. Van Bommel, J.E. Crombeen, A. Van Tooren, LEED and Auger electron observations of the SiC(0001) surface, Surface Science, 48 (1975) 463-472. [30] L.M. Zhang, X.H. Huang, X.L. Song, Fundamentals of Materials Science, Wuhan University of Technology Press, Wuhan, 2004. [31] G. Zhenduo, Z. Zhongtai, J. Jinsheng, Physical Properties of Inorganic Materials, Tsinghua University Press, Beijing, 1992. [32] W. Yang, H. Araki, A. Kohyama, S. Thaveethavorn, H. Suzuki, T. Noda, Process and mechanical properties of in situ silicon carbide-nanowire-reinforced chemical vapor infiltrated silicon carbide/silicon carbide composite, Journal of the American Ceramic Society, 87 (2004) 1720-1725.
18
Fig. 1. Surface SEM images of SiCf/SiC composites (a) before corrosion and after corrosion in liquid FLiNaK salt at (b) 800 °C, (c) 900 °C and (d) 1000 °C for 500 h.
19
Fig. 2. Cross-section SEM images of SiCf/SiC composites (a) before corrosion and after corrosion in liquid FLiNaK salt at (b) 800°C, (c) 900 °C and (d) 1000 °C for 500 h. The insert of (a) is the EDS liner scan result of oxygen element cross the boundary between deposited SiC layers in SiC matrix.
20
Fig. 3. XRD patterns of SiCf/SiC composites (a) before and (b) after corrosion in liquid FLiNaK salt at varying temperatures for 500 h.
21
Fig. 4. Stress - strain curves of SiCf/SiC composites before and after corrosion. Samples corroded in FLiNaK salt at corrosion temperature and ruptured at flexural temperature are marked as “corrosion temperature @ flexural temperature”. The “Blank” is the representation of as-received samples.
22
Fig. 5. Fracture surface SEM images of (a) as-received SiCf/SiC composites ruptured at 800 °C and corroded SiCf/SiC composites in FLiNaK salt (corrosion temperature @ fracture temperature): (b) 800 °C @ 800 °C, (c) 900 °C @ 900 °C and (d) 1000 °C @ 1000 °C.
23
Fig. 6. Illustration of microstructure evolution and crack growth during rupture in SiCf/SiC composites before and after corrosion in FLiNaK salt (corrosion temperature @ fracture temperature): (a)(b) as-received samples, (c)(d) 800 °C @ 800 °C, (e)(f) 900 °C @ 900 °C and (g)(h) 1000 °C @ 1000 °C.
24
Table 1 Mass loss ratio of SiCf/SiC composites after immersion corrosion in liquid FLiNaK salt at various temperature for 500 h. Every bar in each experiment group was weighed 3 times independently to obtain the weight of each bar, then use the data of 5 bars in each experiment group to calculate the average mass loss ratio, mass loss rate and the corresponding standard deviation. Experimental Group
800 °C-500 h
900 °C-500 h
1000 °C-500 h
Mass loss ratio / %
5.914±0.578
15.144±0.642
31.062±0.707
Mass loss rate / μg·mm-2·h-1
0.211±0.021
0.537±0.023
1.107±0.024
25
Table 2 High temperature flexural strength of SiCf/SiC composites before and after corrosion in liquid FLiNaK salt at varying temperatures. Each experiment group tested 3 bars to calculate the average flexural strength and standard deviation. Flexural strength / MPa Flexural temperature / °C
Flexural strength retention ratio / % Before corrosion
After corrosion*
800
450.60 ± 90.49
232.67 ± 44.68
51.64
900
270.86 ± 26.82
176.77 ± 13.33
65.26
1000
238.89 ± 4.63
60.03 ± 11.81
25.13
* The corrosion temperature is as same as the flexural temperature.
26
Table 3 Porosity of SiCf/SiC composites before and after corrosion in liquid FLiNaK salt at varying temperature for 500 h. Each group tested 5 bars to calculate the average porosity and standard deviation Corrosion temperature / °C Sample
Porosity / %
As-received
5.01±0.13
800
900
1000
5.37±0.57
6.73±0.38
12.42±1.88
27